IN THIS CHAPTER
Figuring out why you need a matrix
Computing with matrix calculus to your advantage
Getting a glance at how probability works
Explaining the Bayesian point of view on probability
Describing observations using statistical measures
If you want to implement existing machine learning algorithms from scratch or you need to devise new ones, you require a profound knowledge of probability, linear algebra, linear programming, and multivariable calculus. You also need to know how to translate math into working code, which means having sophisticated computing skills. This chapter begins by helping you understand the mechanics of machine learning math and describes how to translate math basics into usable code.
If you want to apply existing machine learning for practical purposes instead, you can leverage existing R and Python software libraries using a basic knowledge of math and statistics. In the end, you can’t avoid having some of these skills because machine learning has strong roots in both math and statistics, but you don’t need to overdo it. After you get some math basics down, the chapter shows how even simple Bayesian principles can help you perform some interesting machine learning tasks.
Even though this introductory book focuses on machine learning experiments using R and Python, in the text you still find many references to vectors, matrices, variables, probabilities, and their distributions. The book ...